Heuristic #9
Heuristic #9
Conversation
There was a problem hiding this comment.
I guess this belongs to the RL agent and therefore shouldn't be part of this branch / PR
| from ..bin_schedule import _do_bin_pack | ||
|
|
||
|
|
||
| class Option(Protocol): |
There was a problem hiding this comment.
Rename to PartitioningScheme or InitialPartitioningScheme
| """ | ||
|
|
||
| schedules = [] | ||
| num_cores = max(len(OPTIONS), cpu_count()) |
There was a problem hiding this comment.
Should this be a min? Otherwise, using more cores than physically available has probably rather negative impacts on performance.
| if "partition_size" not in kwargs: | ||
| partition_size = 10 | ||
| else: | ||
| partition_size = kwargs["partition_size"] |
There was a problem hiding this comment.
| if "partition_size" not in kwargs: | |
| partition_size = 10 | |
| else: | |
| partition_size = kwargs["partition_size"] | |
| partition_size = kwargs.pop("partition_size", 10) |
or equivalently with get if you don't want to mutate the kwargs
| if partition[-1] == 1: | ||
| partition[-1] = 2 | ||
| partition[-2] -= 1 |
There was a problem hiding this comment.
Maybe add an assertion that partition[-2] > 2 before subtracting from it
| def is_feasible(schedule: Schedule) -> bool: | ||
| """Checks if a schedule is feasible.""" | ||
| return all( | ||
| sum(job.circuit.num_qubits for job in bucket.jobs) <= machine.capacity | ||
| for machine in schedule.machines | ||
| for bucket in machine.buckets | ||
| ) |
There was a problem hiding this comment.
Why not implement this in the Schedule class?
| ] | ||
| benchmark_results: list[Result] = [] | ||
| for benchmark in benchmarks: | ||
| problme_circuits = _cut_circuits(benchmark, setting) |
| return results | ||
|
|
||
|
|
||
| def _cut_circuits( |
There was a problem hiding this comment.
There is significant code duplication (_cut_circuits, _generate_partitions and _partition_big_to_small) are similar in initialize.py
| for machine in schedule.machines: | ||
| if len(machine.buckets) == 0: | ||
| continue | ||
| swap_buckets = np.random.randint(1, dtype=bool) |
There was a problem hiding this comment.
Does it make sense that all swaps are done either on bucket or on job level? In other words, wouldn't it make sense to randomly assign swap_buckets again in each iteration of for _ in range(number_of_swaps)?
|
|
||
| def _diversify(population: list[Schedule]) -> list[Schedule]: | ||
| """Swaps jobs between different machines.""" | ||
| local_population = population.copy() |
There was a problem hiding this comment.
local_population is not used. Instead, you are editing the passed population (which still kind of works because local_population is only a shallow copy)
I've added a scatter search heuristic for finding a schedule.
Todos:
For more todos, see
generate_heuristic_info_schedule